Abacavir is associated with increased risk of cardiovascular disease in HIV-infected patients: A UK clinic case-control study Collins Iwuji, Duncan Churchill, Yvonne Gilleece, Martin Fisher Brighton and Sussex University Hospital, NHS Trust Introduction Abacavir associated with increased cardiovascular risk Study Design Event ascertainment n Effect of ABC D:A:D A RS Prospective Case control in Prospective, pre-defined 33,347 Yes Prospectively reported MI, retrospectively validated 289 cases 884 controls Yes 1 st yr exposure SMART RCT observational Prospective pre-defined 2,752 Yes Quebec (QPHID) Case control in STEAL RCT Prospective Retrospective, MI identified via ICD-9 disease codes 125 cases 1,084 controls 357 Yes Yes Danish cohort Prospective Prospective, MI data from Danish National Hospital Registry 2,952 Yes D:A:D study group. Lancet 2008;371:1417-26 Lang S et al. ANRS CO4. 16 th CROI, 8-11 Feb 2009, Montreal, Canada. LB43 Lundgren et al. AIDS 2008;22:F17-F24 Durand M et al. 5 th IAS 19-22 ND July 2009 Cape Town. Poster TUPEB175 Cooper et al. 16 th CROI, 8-11 Feb 2009; Montreal, Canada. Abstract 576 Obel N et al. HIV Medicine 2010; 11:130-136
No association between abacavir and increased cardiovascular risk Study Design Event ascertainment n Effect of ABC GSK analysis ALLRT ACTG 5001 Veterans Administration BICOMBO HEAT 54 trials (12 RCTs) Retrospective database search LTFU from 5 RCTs Retrospective by 2 independent reviewers Retrospective Retrospective substudy of RCT RCT Retrospective, AMI (& CVA) identified via ICD-9 disease codes Assessment of CV biomarkers Assessment of CV biomarkers 14,174 No 3,205 No 19,424 No 80 No 688 No Brothers C et al. JAIDS 2009;51:20-28 Benson C et al. 16 th CROI, 8-11 Feb 2009, Montreal. Abstrsct 721 Bedimo et al. IAS 19-22 July 2009 Cape Town. Abstract MOAB202 Martinez et al. IAS 19-22 July 2009 Cape Town. Abstract MOAB203 McComsey G et al. 16 th CROI, 8-11 Feb 2009, Montreal. Abstract 732 Objectives Case-control study to assess the association between the risk of cardiovascular disease (CVD) and exposure to abacavir. Cumulative exposure to protease inhibitors (PI) Cumulative exposure to NNRTI class
Cases 3440 patients had at least one appointment in the clinic between Oct 1994 and Oct 2009 First diagnosis of cardiovascular disease (MI, strokes or coronary artery disease - confirmed) within the clinic cohort between Oct 1994 and Oct 2009 51 cases were identified, 13 cases were excluded 4 cases of MI occurred before the diagnosis of HIV 2 cases of MI occurred before the study period 1 case of MI, 3 cases of strokes and 3 cases of coronary artery disease did not have enough medical records Controls 124 HIV-infected patients with no history of CVD Matched with corresponding CVD case for duration in follow up, age and sex Up to 5 matched controls selected for each case without replacement Cases eligible as control up to the time of diagnosis of the cardiovascular disease
Methods Data collected for cases and control Cardiovascular risk factors Smoking, dyslipidaemia, hypertension, family history, diabetes Treatment for dyslipidaemia, hypertension and diabetes Estimated glomerular filtration rate (egfr) HIV related factors CD4 count within 3 months of CVD Plasma HIV-1 RNA within 3 months of CVD Antiretroviral therapy history Statistical analysis Conditional logistic regression models were constructed to assess the effect of Model 1: cumulative exposure to abacavir Model 2: Recent and cumulative exposure to abacavir Model 3: Recent, cumulative and Past exposure to abacavir Potential confounders which affected the association between abacavir and the risk of CVD by at least 10% in any of the models were included from Age, smoking, family history, hypertension, diabetes, dyslipidaemia Estimated glomerular filtration rate (egfr) CD4 count and plasma HIV RNA within 3 months of CVD Cumulative use of PI class, NNRTI class and other NRTIs Unadjusted Odd ratios (OR) for cumulative use of individual NRTI, NNRTI and PI class reported
Demographic characteristics Cases (n = 38) Controls (n= 124) Sex, male, n (%) 38 (100) 124 (100) Age, years, median (IQR) 51.4 (43.9-66.6) 50.4 (38.5-69.6) Sexuality Homosexual 35 (92.1) 111 (89.5) Heterosexual 2 (5.3) 11 (8.9) Unknown 1 (2.6) 2 (1.6) egfr, mean (SD) 85(20.0) 88.6(19.3) CD4 count, cells/mm3, within 3 months before event, median (IQR) Viral load < 50 copies/ml, within 3 months before event n (%) Number of CV risk factors, n (%) 492 (315-780) 482 (303-629) 28/35 (80) 92/111(82.9) 0 2 (5.3) 26 (21.0) 1-2 20 (52.6) 87 (70.2) 3 16 (42.1) 11 (8.9) Clinical characteristics Cases (n = 38) Controls (n = 124) Hypertension, n (%) 16 (42.1) 13 (10.5) Diabetes, n (%) 8 (21.1) 6 (4.8) Dyslipidaemia, n (%) No 16 (42.1) 88 (71.0) Yes 22 (57.9) 33 (26.6) Unknown - 3 (2.4) Smoking, n (%) Never 5 (13.2) 36 (29.0) Current 26 (68.4) 63 (50.8) Ex-smoker 5 (13.2) 17 (13.7) Unknown 2 (5.3) 8 (6.5) Family history, n (%) No 5 (13.2) 62 (50) Yes 15 (39.5) 19 (15.3) Unknown 18 (47.4) 43 (34.7)
P=0.01 P=0.73 P=0.26 P=0.13 P=0.85 P=0.25 P=0.69 P=0.33 P=0.33 P=0.46 Exposure to abacavir and risk of CVD Model 1: cumulative exposure only Model 2: cumulative and recent exposure Model 3: cumulative, recent and past exposure Cumulative exposure (per year) 1.15 (0.92-1.43) 0.81 (0.58-1.12) p=0.20 0.80 (0.58-1.12) p=0.19 Recent exposure - 15.81 (2.53-98.79) p=0.003 13.91 (2.08-92.90) p=0.007 Past exposure - - 0.22 (0.02-2.52) Data are OR (95% CI) after adjusting for factors satisfying criteria for confounding - smoking status, family history of CVD, hypertension, diabetes, dyslipidaemia, egfr. No interaction between recent exposure to abacavir and the number of CVD risk factors on the risk of CVD p = 0.15
Exposure to abacavir and risk of CVD Model 1: cumulative exposure only Model 2: cumulative and recent exposure Model 3: cumulative, recent and past exposure Cumulative exposure (per year) 1.15 (0.92-1.43) 0.81 (0.58-1.12) p=0.20 0.80 (0.58-1.12) p=0.19 Recent exposure - 15.81 (2.53-98.79) p=0.003 13.91 (2.08-92.90) p=0.007 Past exposure - - 0.22 (0.02-2.52) Data are OR (95% CI) after adjusting for factors satisfying criteria for confounding - smoking status, family history of CVD, hypertension, diabetes, dyslipidaemia, egfr, and cumulative exposure to abacavir No interaction between recent exposure to abacavir and the number of CVD risk factors on the risk of CVD p = 0.15 Exposure to abacavir and risk of CVD Model 1: cumulative exposure only Model 2: cumulative and recent exposure Model 3: cumulative, recent and past exposure Cumulative exposure (per year) 1.15 (0.92-1.43) 0.81 (0.58-1.12) p=0.20 0.80 (0.58-1.12) p=0.19 Recent exposure - 15.81 (2.53-98.79) p=0.003 13.91 (2.08-92.90) p=0.007 Past exposure - - 0.22 (0.02-2.52) Data are OR (95% CI) after adjusting for factors satisfying criteria for confounding - smoking status, family history of CVD, hypertension, diabetes, dyslipidaemia, egfr, and cumulative exposure to abacavir No interaction between recent exposure to abacavir and the number of CVD risk factors on the risk of CVD p = 0.15
Discussion Similar findings to the D.A.D study Recent exposure to abacavir associated with increased risk of CVD No association with cumulative or past use after adjustment for recent use Discussion Limitations Few events/small sample size, hence composite case definition Retrospective analysis of case note entries Consistency in prospective studies suggests association may not be due to chance Continuing work needed to clarify association between abacavir and CVD
Acknowledgements Our thanks to Prof. Caroline Sabin Stuart Tilbury Dr Andy Skingsley Dr Jenny Whetham